The present invention generally relates to data recording devices such as magnetic disk devices, optical disk devices, and magnetic tape devices, and, more particularly, to a data recording device that rearranges and records data strings on a recording medium that can be set to a data reproducing device that reproduces recorded data in accordance with likelihood information generated from a reproduction signal.
The present invention also relates to a data reproducing device that is suitable for reproducing data from a recording medium on which data is recorded by the above data recording device.
A method of rearranging (interleaving) data in an encoding process has been employed as a technique for preventing burst errors in restored signals in a decoding process. A turbo encoding process and a repetitive decoding process have been used as an encoding and decoding technique respectively using the data rearranging method.
Turbo encoding is an encoding technique with a great encoding gain, and has been widely used in the field of communications. A turbo encoding device generally has two recursive structure convolution encoders for encoding a data bit string u, as shown in FIGS. 1 and 2.
In FIG. 1, the turbo encoding device includes a first encoder 11, an interleaver (π1) 12, a second encoder 13, and a combining unit 14. The data bit string u is supplied to the first encoder 11 and the second encoder 13 via the interleaver (π1) 12.
The first encoder 11 and the second encoder 13 are recursive structure convolution encoders. The first encoder 11 generates a parity bit string p1 from the supplied data bit string u. The interleaver (π1) 12 outputs a signal string having a different bit order from the inputted data bit string u. The second encoder 13 generates a parity bit string p2 from the signal string supplied from the interleaver (π1) 12.
The combining unit 14 combines the data bit string u, the parity bit string p1 outputted from the first encoder 11, and the parity bit string p2 outputted from the second encoder 13 by predetermined rules, to generate an encoded data bit string yk. In the process of combining the data bit string u, the parity bit string p1, and the parity bit string p2, the combining unit 14 removes certain bits (as a puncture function) to increase the encoding ratio. The encoded data bit string yk generated in this manner is outputted from the turbo encoding device. In a communication system, the encoded data bit string yk is modulated by predetermined rules, and is transmitted from a transmitter.
The turbo encoding device shown in FIG. 2 has two recursive convolution encoders (a first encoder 11B and a second encoder 13B) connected in series. In this structure, the data bit string u is encoded by the first encoder 11B, and the bit order in the signal string obtained by the encoding is rearranged by the interleaver (π1) 12. The signal string outputted from the interleaver (π1) 12 is encoded by the second encoder 13, and the signal string obtained by the encoding is outputted as the encoded data bit string yk.
After receiving the signal transmitted from the transmitter in the above manner, the receiver decodes the received signal to obtain signal value strings U, Y1, and Y2, respectively corresponding to the data bit string u and the parity bit strings p1 and p2 contained in the encoded data bit string yk. These signal value strings U, Y1, and Y2 are inputted into the decoding device corresponding to the turbo encoding device.
In the decoding device, soft output decoding is carried out by two decoders corresponding to the two encoders 11 and 13 (11B and 13B), and the soft output information (likelihood information) of each information bit obtained from one of the two decoders is supplied as a priori information to the other decoder. This operation is repeated by the decoding device, which has a structure shown in FIG. 3. FIG. 3 shows an example structure of the decoding device that processes the decoding signal value strings U, Y1, and Y2 corresponding to the data bit string u and the parity bit strings p1 and p2 contained in the encoded data bit string yk outputted from the turbo encoding device shown in FIG. 1.
In FIG. 3, the decoding device includes a first soft input-output decoder (SISO: Soft In Soft Out) 21, interleavers (π1) 22 and 23, a deinterleaver (π1−1) 25, a second soft input-output decoder (SISO) 24, and a hard reference unit 26. The first soft input-output decoder 21 corresponds to the first encoder 11, and the second soft input-output decoder 24 corresponds to the second encoder 13.
The first soft input-output decoder 21 receives the signal value strings U and Y1 and the a priori information L(u) supplied from the second soft input-output decoder 24. The first soft input-output decoder 21 then carries out a maximum a posteriori probability (MAP) decoding process to estimate the a posteriori probability of each bit. Here, the a posteriori probability is the probability of a bit uk being 0 or 1 where the signal value string is Y(y0, y1, . . . , y1, . . . , yn). In the MAP decoding process, the log likelihood ratio L(u*) that is the log ratio of the a posteriori probability P (uk|Y) is calculated by the following formula:L(u*)=L(uk|Y)=1n[P(uk=1|Y)/P(uk=0|Y)]  (1)
In this formula (1), the signal value string Y represents the signal value strings U and Y1.
The probability P (uk=1|Y) of the bit uk being 1 and the probability P (uk=0|Y) of the bit uk being 0 are calculated based on the Trelis diagram that represents the state transition obtained from the signal value strings U and Y1.
The log likelihood ratio L(u*) can be represented as follows:L(u*)=Lc·yk+L(uk)+Le(uk)  (2)
wherein
Lc·yk represents the communication path value, with Lc representing the constant determined by S/N (the communication path value constant) and yk representing the received signal value string of y0, y1, . . . , yn,
L(uk) represents the a priori information that is the known appearance probability of the bit uk being 1 or 0, and
Le(uk) represents the external likelihood information obtained with respect to the bit uk from code restraint.
From the formula (2), the first soft input-output decoder 21 calculates the external likelihood information Le(uk) by the following formula:Le(uk)=L(u*)−Lc·yk−L(uk)  (3)
The log likelihood ratio L(u*) calculated in the above described manner (by the formula (1)) is assigned to the formula (3) to obtain the external likelihood information Le(uk). The string of the external likelihood information Le(uk) obtained in this manner is supplied as the string of the a priori information L(uk) to the second soft input-output decoder 24 via the interleaver (π1) 23. The second soft input-output decoder 24 also receives the signal value string Y2 as well as the signal value string U via the interleaver (π1) 22.
Taking the a priori information L(uk) into consideration, the second soft input-output decoder 24 calculates a new log likelihood ratio L(u*) by the formula (1). Using the new log likelihood ratio L(u*) and the a priori information L(uk) supplied from the first soft input-output decoder 21, the second soft input-output decoder 24 then calculates the external likelihood information Le(uk) by the formula (3).
The external likelihood information Le(uk) obtained by the second soft input-output decoder 24 is then supplied as the a priori information L(uk) to the first soft input-output decoder 21 via the deinterleaver (π1−1) 25. Taking the a priori information L(uk) into consideration, the first soft input-output decoder 21 calculates the log likelihood ratio L(u*) and the external likelihood information Le(uk) in the above described manner. The external likelihood information Le(uk) obtained here is used as the a priori information L(uk) for the second soft input-output decoder 24.
In the above described manner, the first soft input-output decoder 21 and the second soft input-output decoder 24 repeat the process of calculating the log likelihood ratio L(u*), using the external likelihood information Le(uk) calculated by each other decoder as the a priori information L(uk). This process is referred to as a repetitive decoding process. In the first process carried out by the first soft input-output decoder 21, the a priori information L(uk) is set at zero (L(uk)=0).
The hard reference unit 26 determines whether the bit uk is 0 or 1, based on the log likelihood ratio L(u*) obtained by the second soft input-output decoder 24 after the decoding process has been repeated a predetermined number of times. If the log likelihood ratio L(u*) is positive (L(u*)>0), the hard reference unit 26 determines that the bit uk is 1 (uk=1). If the log likelihood ratio L(u*) is negative (L(u*)<0), the hard reference unit 26 determines that the bit uk is 0 (uk=0). The hard reference result is then outputted as a decoding result Uk.
In the above repetitive decoding process, the probability of the bit uk being an originally expected value (1 or 0) becomes greater, and the probability of the bit uk being an unexpected value becomes smaller (i.e., the difference between the probability of the bit uk being 1 and the probability of the bit uk being 0 becomes greater). The reliability of the reference by the hard reference unit 26 increases accordingly.
It is being considered that the turbo encoding and decoding method used in the communication system as described above should be applied to a data recording and reproducing device such as a magnetic disk device or an optical disk device. Examples of application of the turbo encoding and decoding method to a magnetic disk device are found in “W. E. Ryan, ‘Performance of High Rate Turbo Codes on a PR4-Equalized Magnetic Recording Channel’, Proc. IEEE Int. Conf. On Communications, pp 947-951, 1998”.
In such a data recording and reproducing device, the above described turbo encoding method is employed in the recording system (the write system) for writing data onto a recording medium, and the above described repetitive decoding method is employed in the reproducing system (the read system) for reproducing the data from the recording medium. By employing these methods, the high-density data recording on the recording medium (such as a magnetic disk, an optical disk, an magneto-optical disk, or magnetic tape) can be reproduced with few errors.
The problem with a data reproducing device is generally the noise caused at the time of decoding data. The recording and reproducing properties of a data recording and reproducing device such as an optical disk device contain low-pass components, and a high-pass filter is normally used to restrain the noise caused by the low-pass components. A recording data string has a random bit order, and the cut-off frequency of the high-pass filter cannot be increased if the signal components of the recording string contain low-pass components. For this reason, the noise caused in the data recording and reproducing device cannot be effectively restrained.
To solve this problem, there is a known method for restraining the low-pass components in recording data by modulating the recording data or convoluting a predetermined bit string (or a parity bit string). According to this method, the low-pass components of a signal to be actually written onto a recording medium can be restrained by the data modulation or the convolution of a parity bit string, even of the recording data itself contains the low-pass components.
It is possible to apply the technique of data modulation and parity bit string convolution for removing the low-pass components to a data recording and reproducing device that rearranges the bit order in data in the data recording process and generates likelihood information (such as a log likelihood ratio) from a reproduction signal in the data reproducing process, like the above described data recording and reproducing device employing the turbo encoding method and the repetitive decoding method.
In such a data recording and reproducing device, however, the restraint of the low-pass components is not always guaranteed due to the rearrangement of the bit order, if the data modulation or the parity bit string convolution is carried out prior to the bit order rearrangement. On the other hand, if the data modulation or the parity bit string convolution is carried out after the data bit order rearrangement, accurate likelihood information corresponding to the actual recording data cannot be obtained in the data reproducing process.
Specific references to the prior art relating to the present invention include the following:                1) Japanese Laid-Open Patent Application No. 8-287617        
This reference teaches a method of uniformly scattering information words after a process of rearranging the order of the information words in block data. According to this method, adverse influence due to a burst error in a data reproducing process can be reduced.
However, this reference does not teach a method for restraining the low-pass components in the data to be recorded.                2) Japanese Laid-Open Patent Application No. 3-286624        
This reference teaches a method of determining such a control code pattern as to restrain the low-pass components in the signal components in data when an error detection and correction code is added to a data group containing information data and control data.
However, this method does not involve interleaving (or rearrangement) of the data.                3) Japanese Laid-Open Patent Application No. 6-176495        
This reference teaches a method of inserting a bit into every unit word in data so that desired frequency characteristics can be obtained in the process of generating a recording signal.
However, this method does not involve interleaving (or rearrangement) of the data either.                4) Japanese Laid-Open Patent Application No. 62-030436        
This reference teaches a method of restraining the low-pass components of the signal components in data by selecting a suitable position in information data to which an error detection and correction code is added.
However, this method does not involve interleaving (or rearrangement) of the data either.